AI Ethics Disclosure

1. Transparency in AI Functionality

How Our AI Works:

  • The Intelligence Suite (REAL Connections, CrossLink, UDL Pathways, EdConnect) uses generative AI to:

    • Suggest standards-aligned, real-world applications (REAL Connections)

    • Identify interdisciplinary links between subjects (CrossLink)

    • Provide inclusive lesson adaptations (UDL Pathways)

    • Generate instructional design recommendations (EdConnect)

  • AI Models Used: Our tools are built on DeepSeek-V3-0324 with custom educational fine-tuning

  • Human Oversight: No automated decision-making; all outputs require educator review/editing

  • Limitations: We clearly communicate what our AI can and cannot do (e.g., cannot replace professional judgment)

Data Sources:

  • Curated educational frameworks (e.g., ISTE Standards, UDL Guidelines, state/national standards)

  • Peer-reviewed educational research and best practices

  • No student personal data used to train models

  • Ethically sourced content with proper attribution

2. Bias Mitigation

Proactive Measures:

  • Regular audits of all tools for cultural/gender representation and inclusive language

  • Diverse development team with backgrounds in education, psychology, and instructional design

  • UDL Pathways includes counter-bias prompts (e.g., "Ensure case studies represent diverse cultures")

  • REAL Connections' community partner suggestions are vetted for equity (e.g., rural/urban balance, socioeconomic diversity)

  • Continuous monitoring of outputs for emergent bias patterns

User Reporting:

  • Educators may flag biased outputs via ethics@connectedclassroom.org for immediate review

  • Anonymous reporting option available

  • Quarterly bias mitigation reports shared with user community

  • Dedicated Ethics Review Board with diverse representation reviews all significant concerns

3. Human-in-the-Loop Design

Educator Control:

  • AI-generated content is always suggestive, not prescriptive

  • "Override" options in all tools (e.g., manually edit CrossLink's subject pairings)

  • Customization settings to align with local educational priorities and contexts

  • Progressive disclosure of AI capabilities to prevent over-reliance

  • Clear indication of AI-generated vs. human-curated content

Professional Judgment:

  • Tools explicitly prompt educators to apply their professional judgment

  • Training resources on effective AI collaboration for educators

  • Regular user feedback collection to improve human-AI collaboration

4. Privacy by Design

Minimal Data Use:

  • AI processes inputs (e.g., lesson plans) in temporary sessions without retaining context

  • Data segregation between AI processing and user management systems

  • Regular privacy impact assessments conducted by independent experts

5. Accountability

Redress:

  • If AI tools cause harm (e.g., inappropriate suggestions), we commit to:

    • Investigate within 72 hours

    • Remove offending patterns from training data

    • Implement corrective measures

    • Document lessons learned and preventive measures

Governance:

  • Ethics Advisory Board with educators, founders, and community representatives

  • Regular ethics training for all staff members

  • Whistleblower protections for employees reporting ethics concerns

6. Global Collaboration Ethics

For Create, Connect, Compete challenges:

  • Cultural Sensitivity: AI moderates peer feedback to prevent harmful language

  • Attribution: Student work reused (e.g., as examples) always includes creator credit

  • Inclusive Design: Competition prompts designed to be accessible across cultural contexts

  • Digital Citizenship: Clear guidelines for respectful global collaboration

  • Equitable Access: Accommodations for varying technology access levels

Why This Matters for Educators

Aligns with:

  • ISTE Standards for Educators (2.5a: "Design authentic learning activities that include AI-powered tools")

  • Department of Education AI Recommendations (2023) on bias/privacy in EdTech

  • UNESCO Recommendation on the Ethics of Artificial Intelligence (2021)

  • Student Data Privacy Consortium (SDPC) guidelines

Supports Educational Goals:

  • Promotes digital citizenship and critical thinking about technology

  • Models responsible innovation for students

  • Ensures technology serves educational equity rather than undermining it

  • Protects the professional autonomy and judgment of educators

Our Ongoing Commitment: We view this ethics disclosure not as a static document but as an evolving commitment. We welcome educator feedback and regularly update our practices as AI technology and educational needs evolve. Our mission is to ensure AI serves as a tool for educational empowerment, equity, and excellence.